Morphology of the Structural Root System of Sitka Spruce 2. Computer Simulation of Rooting Patterns

Abstract
In a previous paper (Henderson et al . 1983), it was suggested that the processes of root growth and development proceed with a geometric regularity and that, as a consequence, root distribution is further extended and spatially more even than if growth were at random. We examine this hypothesis and the relative importance of the component growth processes through computer simulation of a model for root distribution at a fixed time. Root segments were measured on 16 year trees and statistical distributions fitted to the occurrence of lengths, branching frequencies and growth directions. These distributions comprise the model which assumes that a root system consists of a number of first-order roots originating at the stem, a number of second-order roots originating on first-order ones and so on. Each root includes a number of bends and lateral branching points and terminates in either a fork or when diameter reaches 5 mm, smaller roots not being included. Parameter manipulation of the fitted distributions and further simulation showed that some regular growth mechanisms were necessary for the simulation of realistic rooting patterns. In particular it was important that direction changes at bends and of new roots at forks were typically small to ensure that the system spread outwards, away from the stem. Lateral branches needed to subtend large angles to their parents in order to exploit separate soil regions. Other necessary rules were that first-order roots were almost regularly distributed around the stem and for a tendency for azimuth changes at bends to be alternately clockwise then anticlockwise. Simulations were also used to examine the possibility of estimating total root length from a study of only part of a root system. An example of excavating one quarter of the system is considered and the results indicate that root systems may be so variable that no reliable estimate can be obtained.